Multiple Linear Regression and Large Scale Integration Technology Application to the Texas Instrument TI-59

Abstract

This program takes ordered strings of data and 1)
creates a raw-score sum-of-squares-and-cross-products
matrix (SSCP), 2) provides a fast, efficient method of
constructing a correlation matrix from the SSCP matrix,
3) computes means and standard deviations for all variables,
4) calculates slopes and intercepts for any combination
of two vectors taken as a X-Y pair, 5) provides a least squares
solution for a multiple regression analysis,
where predictor variables (N = 1-6) are regressed on one
criterion, then computes regression coefficients (raw-score)
for these predictors, and 6) computes R2, the variance in
the criterion accounted for by the linear combination of
predictors. Input data may be any combination of discrete
or continuous variables.